Иeureal is the first decentralized open-source protocol designed to produce a distributed artificial intelligence (AI) architecture that is incentivized, maintained and housed within a cryptocurrency.
The project is an infrastructure for predictive algorithms, like an ongoing kaggle. Creator Wil Brown started the project with the idea that AI could be hugely impacted with the help of the technology behind Bitcoin, to pursue his “mad scientist” dream of creating an AI system.
“[My mission is] to bring AI to the people and out of the hands of the few and to create an infrastructure with the potential to be the world’s largest repository of true understanding.”
Brown — along with colleagues Jordan Miller, Brian Bowles and Melanie Swan from Singularity University — decided to enter the Texas Bitcoin Conference hackathon. The team won the top prize in the event from the Dapps Fund, a US$500,000 commitment for purchase from a successful crowdsale. The team has since grown larger and now meets and discusses weekly.
Cointelegraph got together with Brown to discuss the project and Brown’s vision.
“These can give us an intimate view of the inner workings of our own minds. When we peer at our own subconscious, for instance during dreams, we see very similar things. It’s telling that these images are coming from the “subconscious” of working AI algorithms.”
- Wil Brown (left)
Cointelegraph: What is Иeureal and how does it introduce blockchain tech into AI?
Wil Brown: Иeureal currently uses a blockchain to provide a real-time, ongoing and crowdfunded competition platform for predictive AI algorithms.
CT: When did you come up with this project and why?
WB: January 2014, at the height of the bitcoin price, I was organizing the Utah Bitcoin Community and everyone was excited about getting bitcoin projects going. I had a few projects that I was working on, but I realized that this was a chance to pursue my real “mad scientist” dream of creating a system where AI could really grow.
CT: You won the Texas Bitcoin Conference ‘Hackathon.’ Can you tell us about this?
WB: I had just excitedly contacted Melanie Swan, a quantitative methods and prediction markets instructor at Singularity University, about an article she had written describing something very similar to what I was working on with Иeureal. She mentioned that she was giving a talk at the conference and suggested that I enter the hackathon with my project. She joined me and Brian Bowles, and we won the top prize in the hackathon from the Dapps Fund! A US$500,000 commitment for purchase from a successful crowdsale, the same prize that StorJ won last year.
CT: How many people are working with you on the project?
WB: Our core team is Jordan Miller, Brian Bowles and Dyna Elizondo. We have about 16 advisors, including David Johnston, Melanie Swan and Brian Nelson. We are looking to grow our community! Please join us if you are interested in being part of the future of AI.
CT: What do you think of the recent algorithm paintings?
WB: These can give us an intimate view of the inner workings of our own minds. When we peer at our own subconscious, for instance during dreams, we see very similar things. It’s telling that these images are coming from the “subconscious” of working AI algorithms.
“[AI] is a new knowledge revolution that builds upon the information revolution of the Internet. AI is currently providing those with the ability to use it automatic insight into the massive amount of information that surrounds us.”
CT: What do you see the blockchain doing to advance AI?
WB: Just like bitcoin gave an incentive to create the world's largest supercomputer devoted to one purpose, integrating money into a protocol gives it a level of adoptive power that a traditional open source project doesn’t have. And the decentralized nature of blockchain technology gives the power to create and control the AI algorithms to everyone. With Иeureal, a genius child in Iceland can compete directly with an MIT professor with the same amount of effort.
CT: How is AI advancing the human race?
WB: It is a new knowledge revolution that builds upon the information revolution of the internet. AI is currently providing those with the ability to use it automatic insight into the massive amount of information that surrounds us. They are able to do things like cheaply and instantly create personalized re-routes for 870,000 travelers going to many different destinations and traveling on bus or car or train when the Putney Bridge in London is closed for emergency repairs. With the Internet moving ever deeper into everything we use, these people will have unprecedented power over the world.
CT: Do you think that AI and humans will ever have conflicts over ‘rights’?
WB: I don’t know. You should ask the AI when you get the chance. If we try to model AI to be like us, then it will be like us. But there is so much more than being like us that it could be. I don’t think any of us have any real clue what that might be like. Does a dog understand completely what it is like to be a human?
“Those of us that are embracing and learning to become integrated with machines will be, along with AI, the ‘singularity’ of the future. Most of us are already on that path. Today, a football player with a smartphone is integrated more than a cyberpunk of the 80s.”
CT: What's your opinion on singularity as far as cyborgs or cyberpunk is concerned. Do you feel the human race is approaching ideas of this nature?
WB: Yes. Those of us that are embracing and learning to become integrated with machines will be, along with AI, the “singularity” of the future. Most of us are already on that path. Today, a football player with a smartphone is integrated more than a cyberpunk of the 80s. I plan to create a virtual reality interface for Иeureal that will let us literally dance with the AI in a natural, human way. Learning from it and training it at the same time to “become one” with the machine. Like two brain hemispheres forming a corpus callosum. Can you imagine connecting with the whole world in such an intimate way that your best intention flows through the internet like your mind soars when you daydream?
CT: Can you give our readers a description of the marketplace?
WB: The Иeureal marketplace will be like an “app store” where people can buy and sell AI algorithms and the data needed for them. And it will be where people can sponsor public data sources that they want predicted, like the weather in Salt Lake City. It will also provide the mechanism that supports private data prediction and payment.
CT: Can you explain the Иeureal architecture?
WB: The main architecture goal for Иeureal is to ride on top of the best developing blockchain technologies being created. For our prototype, we are using a Peercoin base with proof of work, proof of stake and a new “coinbase” transaction we call the “Prediction Reward Transaction.” The payout of this transaction is calculated by all nodes using data “ticks” from Data Oracles (a public data source) and all the corresponding prediction and sponsorship transactions. These two custom transactions are “coinage” transactions that use up coinage as their opportunity cost. The “Datasource Sponsorship Transaction” lets you spend coinage to sponsor a Data Oracle. The “Miner Prediction Entry” is the transaction that predictors use to submit their predictions to the blockchain. We use a separate UI client built in java and we are making an RPC interface for the AI code so that it can be agnostic to language and platform.
“The main architecture goal for Иeureal is to ride on top of the best developing blockchain technologies being created.”
CT: What is the ‘vulture’ problem?
WB: People can try to predict who the best predictors are and follow them to “vulture” some of their winnings by submitting values close to them. It can be minimized by some fancy strategies, but it reduces the incentive for the best predictors to keep predicting.
CT: Funding for AI research has seen big growth. Do you see this continuing?
WB: It has blown up because of the successes of the industry. But it doesn’t really matter for Иeureal where VC puts their cheap cash. I believe the real advances in AI will come from the independent research of individuals and groups, across the world and in unlikely places.
CT: What's the overall end mission of Иeureal?
WB: To bring AI to the people and out of the hands of the few and to create an infrastructure with the potential to be the world’s largest repository of true understanding.